The present invention relates generally to feedthrough filter capacitor EMI filters. More particularly, the present invention relates to a hybrid EMI filter substrate and/or flex cable assembly which embodies embedded shielded flat-through/feedthrough filters and/or energy dissipating circuit elements. This invention is applicable to a wide range of connectors, terminals and/or hermetic seals that support lead wires as they ingress/egress into electronic modules or shielded housings. In particular, the present invention applies to a wide variety of active implantable medical devices (AIMDs).
FIGS. 1-40 provide a background for better understanding the significance and novelty of the present invention.
FIG. 1 illustrates various types of active implantable and external medical devices 100 that are currently in use. FIG. 1 is a wire formed diagram of a generic human body showing a number of implanted medical devices. 100A represents a family of hearing devices which can include the group of cochlear implants, piezoelectric sound bridge transducers and the like. 100B represents a variety of neurostimulators and brain stimulators. Neurostimulators are used to stimulate the Vagus nerve, for example, to treat epilepsy, obesity and depression.
Brain stimulators are pacemaker-like devices and include electrodes implanted deep into the brain for sensing the onset of the seizure and also providing electrical stimulation to brain tissue to prevent the seizure from actually occurring. The lead wires associated with a deep brain stimulator are often placed using real time MRI imaging. 100C shows a cardiac pacemaker which is well-known in the art. 100D includes the family of left ventricular assist devices (LVAD's), and artificial hearts, including the recently introduced artificial heart known as the Abiocor. 100E includes an entire family of drug pumps which can be used for dispensing of insulin, chemotherapy drugs, pain medications and the like. Insulin pumps are evolving from passive devices to ones that have sensors and closed loop systems. That is, real time monitoring of blood sugar levels will occur. These devices tend to be more sensitive to EMI than passive pumps that have no sense circuitry or externally implanted lead wires. 100F includes a variety of bone growth stimulators for rapid healing of fractures. 100G includes urinary incontinence devices. 100H includes the family of pain relief spinal cord stimulators and anti-tremor stimulators. 100H also includes an entire family of other types of neurostimulators used to block pain. 100I includes a family of implantable cardioverter defibrillator (ICD) devices and also includes the family of congestive heart failure devices (CHF). This is also known in the art as cardio resynchronization therapy devices, otherwise known as CRT devices. 100J illustrates an externally worn pack. This pack could be an external insulin pump, an external drug pump, an external neurostimulator or even a ventricular assist device. 100K illustrates the insertion of an external probe or catheter. These probes can be inserted into the femoral artery, for example, or in any other number of locations in the human body. 100L illustrates one of various types of EKG/ECG external skin electrodes which can be placed at various locations. 100M are external EEG electrodes placed on the head.
FIG. 2 is a prior art unipolar discoidal feedthrough capacitor, which has an active internal electrode plate set 102 and a ground electrode plate set 104. The inside diameter termination surface 106 is connected electrically to the active electrode plate set 102. An outside diameter termination surface 108 is both solderable and electrically conductive, and it is connected to the outside diameter of electrode plate sets 104.
FIG. 3 is a cross-section of the discoidal feedthrough capacitor of FIG. 2 shown mounted to a hermetic seal 112 of an active implantable medical device (AIMD). In prior art discoidal feedthrough capacitor devices, the lead wire 114 is continuous. The hermetic seal 112 is attached to, typically, a titanium housing 116, for example, of a cardiac pacemaker. An insulator 118, like alumina ceramic or glass, is disposed within a ferrule 120 and forms a hermetic seal against body fluids. The terminal pin or lead wire 114 extends through the hermetic seal 112, passing through aligned passageways through the insulator 118 and the capacitor 110. A gold braze 122 forms a hermetic seal joint between the terminal pin 114 and the insulator 118. Another gold braze 124 forms a hermetic seal joint between the alumina insulator 118 and the titanium ferrule 120. A laser weld 126 provides a hermetic seal joint between the ferrule 120 and the housing 116. The feedthrough capacitor 110 is shown surface mounted in accordance with U.S. Pat. No. 5,333,095, and has an electrical connection 128 between its inside diameter metallization 106 and hence the active electrode plate set 102 and lead wire 114. There is also an outside diameter electrical connection 130 which connects the capacitor's outside diameter metallization 108 and hence the ground electrodes 104 to the ferrule 120. Feedthrough capacitors are very efficient high frequency devices that have minimal series inductance. This allows them to operate as EMI filters over very broad frequency ranges. Referring once again to FIG. 3, one can see that another way to describe a prior art discoidal feedthrough capacitor 110 is as a three-terminal capacitor. Three-terminal devices generally act as transmission lines. Referring to FIG. 3, one can see that there is a current “I” that passes into lead wire 114. For a prior art AIMD, on the body fluid side there is generally an implanted lead which can undesirably act as an antenna which can pick up energy from environmental emitters. This energy is known as electromagnetic interference (EMI). Cell phones, microwave ovens and the like have all been implicated in causing interference with active implantable medical devices. If this interference enters lead wire 114 at point X (FIG. 3), it is attenuated along its length by the feedthrough capacitor 110. Upon exiting, the undesirable high frequency EMI has been cleaned off of the normal low frequency (LF) circuit current (such as pacemaker pacing pulses or biologic frequency sensors) so that the high frequency EMI has been significantly attenuated. Another way of looking at this is as the high frequency energy passes from terminal 1 to terminal 2 (FIGS. 3 and 4), it is diverted through the feedthrough capacitor 110 to the ground terminal which is also known as the third terminal or terminal 3. The feedthrough capacitor 110 also performs two other important functions: a) its internal electrodes 102 and 104 act as a continuous part of the overall electromagnetic shield housing of the electronic device or module which physically blocks direct entry of high frequency RF energy through the hermetic seal 112 or equivalent opening for lead wire ingress and egress in the otherwise completely shielded housing (such RF energy, if it does penetrate inside the shielded housing can couple to and interfere with sensitive electronic circuitry), and; b) the feedthrough capacitor 110 very effectively shunts undesired high frequency EMI signals off of the lead wires to the overall shield housing where such energy is dissipated in eddy currents resulting in a very small temperature rise.
FIG. 4 is a schematic diagram showing the discoidal feedthrough capacitor 110 previously described in connection with FIGS. 2 and 3. As one can see, it is a three-terminal device consistent with terminals 1, 2 and 3 illustrated in FIG. 3.
FIG. 5 is a quadpolar prior art feedthrough capacitor 132 which is similar in construction to that previously described in FIG. 2 except that it has four through holes.
Throughout this description, functionally equivalent elements will be given the same reference number, irrespective of the embodiment being shown.
FIG. 6 is a cross-section showing the internal electrodes 102, 104 of the capacitor 132 of FIG. 5.
FIG. 7 is a schematic diagram showing the four discrete feedthrough capacitors comprising the quadpolar feedthrough capacitor 132 of FIGS. 5 and 6.
FIG. 8 is an exploded electrode view showing the inner and outer diameter electrodes of the unipolar feedthrough capacitor 110 of FIGS. 2 and 3. One can see the active electrode plates set 102 and the ground electrode plate set 104. Cover layers 134 are put on the top and bottom for added electrical installation and mechanical strength.
FIG. 9 is an exploded view of the interior electrodes of the prior art quadpolar feedthrough capacitor 132 previously illustrated in FIG. 5. As shown in FIG. 9, the active electrode plate sets are shown as 102 and the ground electrode plates are shown as 104. Cover layers 134 serve the same purpose as previously described in connection with FIG. 8.
FIG. 10 illustrates a prior art quadpolar feedthrough capacitor 132 mounted on top of a hermetic insulator 118 wherein a wire bond substrate 136 is attached to the top as shown. Wire bond pads 138, 138′, 138″, 138″′ and 140 are shown for convenient connection to the internal circuitry of the AIMD. This is more thoroughly described in FIGS. 75 and 76 of U.S. Pat. Nos. 7,038,900 and 7,310,216, the contents of which are incorporated herein.
FIG. 11 is a cross-section taken generally from section 11-11 from FIG. 10. In FIG. 11, the internal circuit traces T1 through T4 to the wire bond pads 138-138′″ are shown. Referring back to FIG. 10, there is an additional wire bond pad 140 shown on the left side of the wire bond substrate 136. This is also shown in FIG. 11. This is a ground connection to the outside diameter of the hermetic seal ferrule 120 and provides a convenient connection point for electronic circuits and the like that need a ground attachment point on the inside of the AIMD.
FIG. 12 is a schematic diagram of the prior art wire bond pad quadpolar hermetic feedthrough 132 of FIG. 10.
FIG. 13 is a prior art monolithic ceramic capacitor (MLCC) 142. These are made by the hundreds of millions per day to service the consumer electronics and other markets. Virtually all cell phones and other types of electronic devices have many of these. In FIG. 13, one can see that the MLCC 142 has a body 144 generally consisting of a high dielectric constant ceramic such as barium titanate. It also has solderable termination surfaces 146 and 148 at either end. These termination surfaces 146 and 148 provide a convenient way to make a connection to the internal electrode plates of the MLCC capacitor 142. FIG. 13 can also take the shape and characteristics of a number of other types of capacitor technologies, including rectangular, cylindrical, round, tantalum, aluminum electrolytic, stacked film or any other type of capacitor technology.
FIG. 14 is a sectional view taken from section 14-14 in FIG. 13. The left hand electrode plate set is shown as 150 and the right hand electrode plate set is shown as 152. One can see that the left hand electrode plates 150 are electrically connected to the external metallization surface 146. The opposite electrode plate set (or right hand plate set) 152 is shown connected to the external metallization surface 148. One can see that prior art MLCC and equivalent chip capacitors are also known as two-terminal capacitors. That is, there are only two ways electrical energy can connect to the body of the capacitor. In FIGS. 13 and 14, the first terminal “1” is on the left side and the second terminal “2” is on the right side.
FIG. 15 is an ideal schematic diagram of the prior art MLCC capacitor 142 of FIG. 13.
FIG. 16 is a more realistic schematic diagram showing the fact that the MLCC 142 structure as illustrated in FIG. 13 has series inductance L. This inductive property arises from the fact that it is a two-terminal device and does not act as a transmission line. That is, its lead wires and associated internal electrodes all tend to add series inductance to the capacitor. It is well known to electrical engineers that MLCC capacitors will self-resonate at a particular frequency. FIG. 17 gives the formula for this resonant frequency. There is always a point at which the capacitive reactance as shown in FIG. 16 is equal and opposite to the inductive reactance. It is this point that these two imaginary components cancel each other out. If it weren't for resistive losses, at the resonant frequency the impedance between 146,1 and 148,2 as shown in FIG. 16 would go to zero. However, the resistive losses of the inductor L and the equivalent series resistance of the capacitor C prevent this from happening. This is better understood by referring to FIG. 18.
Shown in FIG. 18 are three curves. An ideal capacitor curve is shown which is very similar to the response of a feedthrough capacitor, such as shown in FIG. 3. One can see that the attenuation goes up fairly linearly with frequency all the way up to very high frequencies even above 10,000 megahertz (MHz). The MLCC curve is for the capacitor of FIG. 13. At low frequencies, in this case below 100 MHz, the MLCC curve tracks very closely to an ideal or a feedthrough capacitor. However, as the MLCC nears its self-resonant frequency (SRF), its attenuation tends to go up dramatically. This is because when one refers back to FIG. 16, the inductive and capacitive reactance elements are tending to cancel each other out. As previously mentioned, if it weren't for its resistive losses at resonance (SRF), the MLCC chip would look like a short circuit, in which ideal case its attenuation would be infinite. This means that if it weren't for these resistive losses, we would have infinite attenuation at the SRF. Instead what we have is a peak of approximately 60 dB as shown. Above resonance, the MLCC capacitor becomes increasingly inductive and the attenuation drops dramatically. This is an undesirable effect and this is why feedthrough capacitors have generally been the preferred choice for use in EMI broadband filters.
FIG. 19 shows three different size MLCC capacitors C1-C3 connected around a unipolar feedthrough pin or lead wire 114. Self-resonant frequency is dependent upon the internal inductance of a capacitor. This was illustrated and described in connection with FIG. 16. One can reduce the amount of inductance by using a physically smaller MLCC capacitor. For example, referring to FIG. 19, one could have one each of what is known in the art as a size 0402, a 0603 and a 0805 MLCC capacitor. This is an EIA designation wherein, for example, 0805 would be 0.080 inches long and 0.050 inches wide. Accordingly, these three MLCC capacitors C1-C3 would have three different resonant frequencies. This is more thoroughly described in U.S. Pat. No. 5,973,907 and U.S. Pat. No. 5,959,336 the contents of which are incorporated herein by reference. FIG. 20 is the schematic diagram for the three MLCC capacitors of FIG. 19.
FIG. 21 shows the attenuation response for the three chip capacitor unipolar hermetic terminal in FIG. 19. These three capacitors C1-C3 are acting in parallel as shown in the schematic diagram of FIG. 20. Referring to FIG. 21, we can see that there are now three resonant peaks representing the self-resonant frequency of each of these individual MLCC capacitors acting together in parallel. Shown for reference is the ideal capacitor response curve previously shown in FIG. 18. The SRF for C1, C2 and C3 are also shown. The physically largest capacitor C1 will have the lowest self-resonant frequency whereas the physically smaller capacitor (C3) will have the highest self-resonant frequency. This is because, in general, the smaller the MLCC capacitor, the lower its internal inductance. Secondary factors that determine the value of the undesirable equivalent series inductance (ESL) of an MLCC capacitor include the number and spacing of internal electrodes, geometry, form factor and circuit board mounting techniques.
Referring once again to FIG. 19, the reason why this approach has never been commonly practiced in the AIMD market is the fact that this is a complicated design and is also costly. Because of the space limitations and reliability implications, packing this many components into such a small place becomes impractical.
FIG. 22 shows a different method of mounting MLCC capacitors, for example, those previously shown in FIG. 19. In the industry, this is known as the tombstone mounting position, which is a highly undesirable thing to do when the capacitor is to be used as an EMI filter or an RF decoupling device (bad mounting and bad form factor). This is because the capacitor's inductive loop area L1 tends to increase. The increased inductive loop area (integral of area bounded under the loop) has the effect of directly raising the inductance L as previously described in connection with FIG. 16. The reason this is undesirable is this particular capacitor will tend to self-resonate at a much lower frequency (and thereby becomes a less effective high frequency device or EMI filter).
FIG. 23 illustrates a more desirable way to mount the MLCC capacitor 142 of FIG. 22. This is a conventional flat surface mount technique, which has a much lower inductive loop area L2 as shown (area bounded under the loop). Accordingly, even though the two capacitors are identical in size and capacitance value, the MLCC capacitor 142 as shown in FIG. 23 will resonate at a much higher frequency before it starts to become undesirably inductive.
FIG. 24 is known in the art as a reverse geometry MLCC capacitor 142′. For comparative purposes, the physical size of the MLCC capacitor illustrated in FIG. 24 is exactly the same dimensions as the MLCC capacitors 142 previously shown in FIGS. 22 and 23. The important thing is the location of the termination surfaces 146′ and 148′. The MLCC capacitor 142′ in FIG. 24 has been terminated along its long sides. Accordingly, its inductive loop area or the area bounded underneath the loop L3 is the smallest of all the loop configurations. Thus, the capacitor 142′ of FIG. 24 will self-resonate at a much higher frequency as compared to the MLCC capacitors 142 shown in FIGS. 22 and 23. A good treatment of this is found in a technical paper entitled, A CAPACITOR′S INDUCTANCE, which was given at the Capacitor and Resistor Technology Symposium in Lisbon, Portugal, Oct. 19-22, 1999. This paper was co-authored by Robert Stevenson and Dr. Gary Ewell of Aerospace Corporation. A related paper was given entitled, A CAPACITOR′S INDUCTANCE: CRITICAL PROPERTY FOR CERTAIN APPLICATIONS and was given by the same authors at the 49th Electronic and Components Technology Conference of the Institute of Electrical and Electronic Engineers held Jun. 1-4, 1999 in San Diego, Calif.
FIG. 25 is the same electrical schematic diagram as previously illustrated in FIG. 16, but additionally showing the equivalent circuit model for an MLCC. Added are resistors IR and ESR. IR is the insulation resistance of the capacitor C. For electronic circuit analysis reasons, this IR resistor can generally be ignored. The reason for this is that it is typical that the value of IR is in excess of 10 Gigaohms (10,000,000,000 ohms). This number is so high compared to the values of the other components of the capacitor circuit model that it can be safely ignored. Also added to the complete schematic model shown in FIG. 25 is the capacitor series resistance (ESR). This is the total ESR including the dielectric loss tangent of the ceramic materials themselves and all ohmic losses and other electrical connections within and external to the capacitor itself. As previously stated, the presence of resistor ESR is why at the self-resonant frequency, the insertion loss does not go to infinity.
FIG. 26 is a prior art chip transient suppression diode 154, such as a transorb or the like.
FIG. 27 is a schematic diagram showing the diode chip 154 of FIG. 26 connected between an active medical device lead wire 114 and circuit ground. The dashed line shown in FIG. 27 illustrates the shielded housing of the AIMD. The reason for diode chip 154 (or multiple diode arrays) is to help protect the sensitive electronic circuits of the AIMD from external high voltage insults. These could be electrostatic discharges or the application to the patient of automatic (high voltage) external defibrillation (AED). AEDs are commonly now found in government buildings, airports, airplanes and the like. It is very important that a pacemaker not be burned out during the application of an AED external defibrillation event. The diode chip 154 shown in FIGS. 26 and 27 basically is typically an avalanche type diode which is also known in the art as a zener diode. In other words, they do not forward bias or short out until a certain voltage threshold is reached. These are also known in the art as transorbs and also have other market names. Such diodes can be back to back and placed in parallel in order to suppress biphasic high voltage AED defibrillation pulses.
FIG. 28 is a prior art inductor chip 156. There are many manufacturers of these. These can either have ferrite elements or be non-ferromagnetic. They come in a variety of sizes, inductance values and voltage ratings.
FIG. 29 is a schematic diagram of the inductor chip 156 of FIG. 28.
Referring to FIG. 30, one can see that an inductor circuit trace 158 is printed or deposited right on top of a prior art MLCC capacitor 142 to form an MLCC-T 160. The advantage here is that low cost MLCC's which have been produced from very high volume commercial capacitor operations could be utilized and the inductor trace 158 could be printed on as a supplemental operation. This forms a parallel inductor (L)-capacitor (C) resonant L-C circuit which creates a very high impedance at its resonant frequency. This is effective for suppressing a single RF frequency, such as that from Magnetic Resonance Imaging (MRI) equipment, or the like. This is more thoroughly described in U.S. Patent Application Publication No. US 2007-0112398 A1, the contents of which are incorporated herein by reference.
FIG. 31 shows yet another way to deposit an inductor shape 158 onto a separate substrate 162 to form a parallel L-C resonant circuit. For example, the substrate 162 could be of alumina ceramic or other suitable circuit board material. This could then be bonded with a thin adhesive layer 164 to a prior art MLCC capacitor 142. The composite MLCC-T structure 160′, including corresponding metallization surfaces 146 and 148 on opposite ends, is illustrated in the electrical schematic diagram of FIG. 34 where it is evident that the structure forms a parallel L and C “tank” or bandstop circuit.
FIG. 32 is an isometric view of a novel composite monolithic ceramic capacitor-parallel resonant tank (MLCC-T) 160″ which forms a bandstop or tank filter 166 in accordance with previously referenced U.S. patent application Ser. No. 11/558,349. Viewed externally, one can see no difference between the MLCC-T 160″ of the present invention and prior art MLCC capacitor 142 as shown in FIG. 13. However, the novel MLCC-T 160″ has an embedded inductor 162 which is connected in parallel across the capacitor between its opposite termination surfaces 146 and 148.
FIG. 33 illustrates an exploded view of the various layers of the novel MLCC-T tank filter 160″ shown in FIG. 32. The novel MLCC tank (MLCC-T) 160″ includes an embedded inductor 162. At low frequencies, the embedded inductor 162 shorts out the capacitor from one end to the other. However, at high frequency, this forms a parallel tank circuit 166 which is again better understood by referring to the schematic diagram in FIG. 34. Referring once again to FIG. 33, one can see that as the capacitor stacks up from the top, we have an area of blank cover sheets 168 followed by one or more embedded inductor layers 162. These inductor traces can have a variety of shapes as further illustrated in FIG. 83 of U.S. Patent Application Publication No. US 2007-0112398 A1. It will be obvious to those skilled in the art that there are a variety of optional shapes that could also be used. Then there are a number of other blank interleafs 170 before one gets to the capacitor electrode plate sets, 150 and 152. One can see the capacitor electrode plate set 150 which connects to the left hand termination 146 and one can also see the capacitor electrode plate set 152 which connects to the right hand termination 148. In FIG. 33, only single electrodes are shown as 150, 152. However, it will be obvious to those skilled in the art that any number of plates “n” could be stacked up to form the capacitance value that is desired. Then bottom blank cover sheets 168 are added to provide insulative and mechanical strength to the overall TANK filter MLCC-T 160″.
After sintering the composite structure at high temperature, the last step, referring back to FIG. 32, is the application of the solderable termination surfaces 146 and 148. These termination surfaces can be a thick film ink, such as palladium silver, glass frit, gold plating, or the like and applied in many processes that are known in the art. Once again, the overall MLCC-T 160″, which is illustrated in FIG. 32, looks identical to a prior art MLCC 142 as shown in FIG. 13. However, embedded within it is a novel parallel inductor structure 162 creating a novel parallel tank or bandstop filter 166 shown in the schematic diagram of FIG. 34.
Referring to schematic drawing FIG. 34, one can see that the inductor L has been placed in parallel with the capacitor C which is all conveniently located within the monolithic structure MLCC-T 160″ shown in FIG. 32.
In FIG. 35 only one pole of a quadpolar feedthrough capacitor 132 is shown, which is better understood by referring to its schematic diagram shown in FIG. 36. One can see that there is a feedthrough capacitor 132 which is also known as a broadband EMI filter shown as C1, C2, C3 and C4 in FIG. 36. In line with each one of these circuits is a parallel resonant bandstop filter MLCC-T 160 to block MRI pulsed RF frequencies or frequencies from similar powerful emitters. The function of these bandstop filters is better understood by referring to the complete description in U.S. Pat. No. 7,363,090, the contents of which are incorporated herein.
Referring once again to FIG. 35, one can see that there is a metallic ferrule 120 which is attached to a hermetic insulator 118 by means of gold braze 124. There are also two lead wires 114 and 114′ as shown. Lead wire 114 is mechanically and hermetically attached to insulator 118 by means of gold braze material 122. The bandstop filter or tank filter MLCC-T 160 is held in place with an insulative spacer plate 172. The feedthrough capacitor 132 is mounted on top as shown. Lead wire 114′ is attached to the other end of the tank filter MLCC-T 160. A capacitor outside diameter metallization 108 connects to the capacitor's internal ground electrodes 104. Electrical connection 126 is made between the capacitor's outside diameter metallization 108 and both the metal of the ferrule 120 and gold braze material 124.
FIG. 37 is a different type of prior art MLCC feedthrough capacitor 142 that is built into a special configuration. It is known in the art by some as a flat-through capacitor (it also has other trade names). It will be referred to herein as a flat-through capacitor 174. At low frequencies, the flat-through capacitor 174 exhibits ideal capacitance behavior versus frequency. That is, its attenuation curve versus frequency is nearly ideal. This is because it is truly a three-terminal device which acts as a transmission line in a manner similar to those of prior art discoidal feedthrough capacitors 110. This is better understood by referring to its internal electrode plate geometry as shown in FIG. 38. Shown is a through or active electrode plate 175 that is sandwiched between two ground electrode plates 178 and 178′. The through or active electrode plate 175 is connected at both ends by termination surfaces 180 and 182. When the capacitor is mounted between circuit trace lands 184 and 186 as shown in FIG. 37, this connects the circuit trace together between points 184 and 186. Referring to the active circuit trace 175 in FIG. 38, one can see that there is a current i1 that enters. If this is a high frequency EMI current, it will be attenuated along its length by the capacitance of the flat-through capacitor and emerge as a much smaller in amplitude EMI signal at terminal 2 as i1′. Similar to discoidal feedthrough capacitors, the flat-through capacitor 174 is also a three-terminal capacitor as illustrated in FIG. 37. The point of current input i1 is terminal 1, the point of circuit current egress i1′ is known as terminal 2 and ground is known as terminal 3. In other words, any RF currents that are flowing down the circuit trace must pass through the electrodes 175 of the capacitor 174. This means that any RF signals are exposed for the full length of the electrode plate 175 between the ground electrodes 178 and the capacitance that is formed between them. This has the effect of making a very novel shape for a three-terminal feedthrough capacitor. One negative to this type of capacitor 174 is that it is not conveniently mountable in such a way that it becomes an integral part of an overall shield. There is always a frequency at which undesirable RF coupling 188 across the device will occur. This usually does not happen until 100 MHz or above. At very high frequencies, such as above 1 GHz, this problem becomes quite serious. Another negative, as compared to prior art discoidal feedthrough capacitors 110 (where the circuit current passes through a robust lead in a feedthrough hole), is that the flat-through capacitor circuit currents must flow through the electrodes of the flat-through capacitor itself (in prior art discoidal/feedthrough capacitors, the only current that flows in the electrodes is high frequency EMI currents). Monolithic ceramic manufacturing limitations on electrode thickness and conductivity means that prior art flat-through capacitors 174 have relatively high series resistance and can only be rated to a few hundred milliamps or a few amps at best (however, an implantable defibrillator must deliver a high voltage pulse of over 20-amps). Prior art MLCC and flat-through electrodes must be kept relatively thin to promote ceramic grain growth through the electrodes in order to keep the capacitor layers from delaminating during manufacturing or worse yet, during subsequent mechanical or thermal shocks which can cause latent failures.
FIG. 39 is the schematic diagram of the prior art flat-through capacitor 174 as illustrated in FIG. 37. Note that its schematic diagram is the same as that for the feedthrough capacitor 110 shown in FIGS. 2 and 3. The difference is that feedthrough capacitors are inherently configured to be mounted as an integral part of an overall shield which precludes the problem of RF coupling (see FIGS. 5-7).
FIG. 40 illustrates the attenuation versus frequency response curve which is shown generically for the flat-through capacitor of FIG. 37. If it weren't for cross-coupling of RF energy, it would perform as an ideal or nearly perfect capacitor would. However, because of this cross-coupling, there is always going to be a certain frequency at which the attenuation starts to parasitically drop off as shown. This drop off is very undesirable in active implantable medical device (AIMD) applications in that there would be less protection against high frequency EMI emitters such as cellular phones and the like. This parasitic drop off in attenuation due to cross-coupling is even a worse problem in military and space applications where EMI filter attenuation requirements of up to 10 or even 18 GHz, is important (implantable medical applications don't generally require filtering much above 3 GHz due to the effective reflection and absorption of human skin of RF energy at frequencies above 3 GHz). Space and military circuits have to operate in the presence of extremely high frequency emitters, such as GHz radars and the like. Accordingly, there is a need for a flat-through type of capacitor that eliminates the problems associated with this parasitic attenuation degradation due to RF cross-coupling across (or outside of around) the capacitor. In addition, there is also a need for flat-through capacitors that can handle much higher circuits through their “through” electrodes. The present invention fulfills these needs and provides other related advantages.